gvisor/pkg/sentry/vfs/file_description_impl_util.go

529 lines
18 KiB
Go
Raw Normal View History

Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
// Copyright 2019 The gVisor Authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package vfs
import (
"bytes"
"io"
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
"gvisor.dev/gvisor/pkg/abi/linux"
"gvisor.dev/gvisor/pkg/context"
"gvisor.dev/gvisor/pkg/errors/linuxerr"
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
"gvisor.dev/gvisor/pkg/sentry/arch"
fslock "gvisor.dev/gvisor/pkg/sentry/fs/lock"
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
"gvisor.dev/gvisor/pkg/sentry/memmap"
"gvisor.dev/gvisor/pkg/sync"
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
"gvisor.dev/gvisor/pkg/syserror"
"gvisor.dev/gvisor/pkg/usermem"
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
"gvisor.dev/gvisor/pkg/waiter"
)
// The following design pattern is strongly recommended for filesystem
// implementations to adapt:
// - Have a local fileDescription struct (containing FileDescription) which
// embeds FileDescriptionDefaultImpl and overrides the default methods
// which are common to all fd implementations for that filesystem like
// StatusFlags, SetStatusFlags, Stat, SetStat, StatFS, etc.
// - This should be embedded in all file description implementations as the
// first field by value.
// - Directory FDs would also embed DirectoryFileDescriptionDefaultImpl.
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
// FileDescriptionDefaultImpl may be embedded by implementations of
// FileDescriptionImpl to obtain implementations of many FileDescriptionImpl
// methods with default behavior analogous to Linux's.
//
// +stateify savable
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
type FileDescriptionDefaultImpl struct{}
// OnClose implements FileDescriptionImpl.OnClose analogously to
// file_operations::flush == NULL in Linux.
func (FileDescriptionDefaultImpl) OnClose(ctx context.Context) error {
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
return nil
}
// StatFS implements FileDescriptionImpl.StatFS analogously to
// super_operations::statfs == NULL in Linux.
func (FileDescriptionDefaultImpl) StatFS(ctx context.Context) (linux.Statfs, error) {
return linux.Statfs{}, syserror.ENOSYS
}
// Allocate implements FileDescriptionImpl.Allocate analogously to
// fallocate called on an invalid type of file in Linux.
//
// Note that directories can rely on this implementation even though they
// should technically return EISDIR. Allocate should never be called for a
// directory, because it requires a writable fd.
func (FileDescriptionDefaultImpl) Allocate(ctx context.Context, mode, offset, length uint64) error {
return linuxerr.ENODEV
}
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
// Readiness implements waiter.Waitable.Readiness analogously to
// file_operations::poll == NULL in Linux.
func (FileDescriptionDefaultImpl) Readiness(mask waiter.EventMask) waiter.EventMask {
// include/linux/poll.h:vfs_poll() => DEFAULT_POLLMASK
return waiter.ReadableEvents | waiter.WritableEvents
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// EventRegister implements waiter.Waitable.EventRegister analogously to
// file_operations::poll == NULL in Linux.
func (FileDescriptionDefaultImpl) EventRegister(e *waiter.Entry, mask waiter.EventMask) {
}
// EventUnregister implements waiter.Waitable.EventUnregister analogously to
// file_operations::poll == NULL in Linux.
func (FileDescriptionDefaultImpl) EventUnregister(e *waiter.Entry) {
}
// PRead implements FileDescriptionImpl.PRead analogously to
// file_operations::read == file_operations::read_iter == NULL in Linux.
func (FileDescriptionDefaultImpl) PRead(ctx context.Context, dst usermem.IOSequence, offset int64, opts ReadOptions) (int64, error) {
return 0, linuxerr.EINVAL
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// Read implements FileDescriptionImpl.Read analogously to
// file_operations::read == file_operations::read_iter == NULL in Linux.
func (FileDescriptionDefaultImpl) Read(ctx context.Context, dst usermem.IOSequence, opts ReadOptions) (int64, error) {
return 0, linuxerr.EINVAL
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// PWrite implements FileDescriptionImpl.PWrite analogously to
// file_operations::write == file_operations::write_iter == NULL in Linux.
func (FileDescriptionDefaultImpl) PWrite(ctx context.Context, src usermem.IOSequence, offset int64, opts WriteOptions) (int64, error) {
return 0, linuxerr.EINVAL
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// Write implements FileDescriptionImpl.Write analogously to
// file_operations::write == file_operations::write_iter == NULL in Linux.
func (FileDescriptionDefaultImpl) Write(ctx context.Context, src usermem.IOSequence, opts WriteOptions) (int64, error) {
return 0, linuxerr.EINVAL
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// IterDirents implements FileDescriptionImpl.IterDirents analogously to
// file_operations::iterate == file_operations::iterate_shared == NULL in
// Linux.
func (FileDescriptionDefaultImpl) IterDirents(ctx context.Context, cb IterDirentsCallback) error {
return linuxerr.ENOTDIR
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// Seek implements FileDescriptionImpl.Seek analogously to
// file_operations::llseek == NULL in Linux.
func (FileDescriptionDefaultImpl) Seek(ctx context.Context, offset int64, whence int32) (int64, error) {
return 0, linuxerr.ESPIPE
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// Sync implements FileDescriptionImpl.Sync analogously to
// file_operations::fsync == NULL in Linux.
func (FileDescriptionDefaultImpl) Sync(ctx context.Context) error {
return linuxerr.EINVAL
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// ConfigureMMap implements FileDescriptionImpl.ConfigureMMap analogously to
// file_operations::mmap == NULL in Linux.
func (FileDescriptionDefaultImpl) ConfigureMMap(ctx context.Context, opts *memmap.MMapOpts) error {
return linuxerr.ENODEV
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// Ioctl implements FileDescriptionImpl.Ioctl analogously to
// file_operations::unlocked_ioctl == NULL in Linux.
func (FileDescriptionDefaultImpl) Ioctl(ctx context.Context, uio usermem.IO, args arch.SyscallArguments) (uintptr, error) {
return 0, linuxerr.ENOTTY
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
}
// ListXattr implements FileDescriptionImpl.ListXattr analogously to
// inode_operations::listxattr == NULL in Linux.
func (FileDescriptionDefaultImpl) ListXattr(ctx context.Context, size uint64) ([]string, error) {
// This isn't exactly accurate; see FileDescription.ListXattr.
return nil, linuxerr.ENOTSUP
}
// GetXattr implements FileDescriptionImpl.GetXattr analogously to
// inode::i_opflags & IOP_XATTR == 0 in Linux.
func (FileDescriptionDefaultImpl) GetXattr(ctx context.Context, opts GetXattrOptions) (string, error) {
return "", linuxerr.ENOTSUP
}
// SetXattr implements FileDescriptionImpl.SetXattr analogously to
// inode::i_opflags & IOP_XATTR == 0 in Linux.
func (FileDescriptionDefaultImpl) SetXattr(ctx context.Context, opts SetXattrOptions) error {
return linuxerr.ENOTSUP
}
// RemoveXattr implements FileDescriptionImpl.RemoveXattr analogously to
// inode::i_opflags & IOP_XATTR == 0 in Linux.
func (FileDescriptionDefaultImpl) RemoveXattr(ctx context.Context, name string) error {
return linuxerr.ENOTSUP
}
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
// DirectoryFileDescriptionDefaultImpl may be embedded by implementations of
// FileDescriptionImpl that always represent directories to obtain
// implementations of non-directory I/O methods that return EISDIR.
//
// +stateify savable
type DirectoryFileDescriptionDefaultImpl struct{}
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
// Allocate implements DirectoryFileDescriptionDefaultImpl.Allocate.
func (DirectoryFileDescriptionDefaultImpl) Allocate(ctx context.Context, mode, offset, length uint64) error {
return syserror.EISDIR
}
Sentry virtual filesystem, v2 Major differences from the current ("v1") sentry VFS: - Path resolution is Filesystem-driven (FilesystemImpl methods call vfs.ResolvingPath methods) rather than VFS-driven (fs package owns a Dirent tree and calls fs.InodeOperations methods to populate it). This drastically improves performance, primarily by reducing overhead from inefficient synchronization and indirection. It also makes it possible to implement remote filesystem protocols that translate FS system calls into single RPCs, rather than having to make (at least) one RPC per path component, significantly reducing the latency of remote filesystems (especially during cold starts and for uncacheable shared filesystems). - Mounts are correctly represented as a separate check based on contextual state (current mount) rather than direct replacement in a fs.Dirent tree. This makes it possible to support (non-recursive) bind mounts and mount namespaces. Included in this CL is fsimpl/memfs, an incomplete in-memory filesystem that exists primarily to demonstrate intended filesystem implementation patterns and for benchmarking: BenchmarkVFS1TmpfsStat/1-6 3000000 497 ns/op BenchmarkVFS1TmpfsStat/2-6 2000000 676 ns/op BenchmarkVFS1TmpfsStat/3-6 2000000 904 ns/op BenchmarkVFS1TmpfsStat/8-6 1000000 1944 ns/op BenchmarkVFS1TmpfsStat/64-6 100000 14067 ns/op BenchmarkVFS1TmpfsStat/100-6 50000 21700 ns/op BenchmarkVFS2MemfsStat/1-6 10000000 197 ns/op BenchmarkVFS2MemfsStat/2-6 5000000 233 ns/op BenchmarkVFS2MemfsStat/3-6 5000000 268 ns/op BenchmarkVFS2MemfsStat/8-6 3000000 477 ns/op BenchmarkVFS2MemfsStat/64-6 500000 2592 ns/op BenchmarkVFS2MemfsStat/100-6 300000 4045 ns/op BenchmarkVFS1TmpfsMountStat/1-6 2000000 679 ns/op BenchmarkVFS1TmpfsMountStat/2-6 2000000 912 ns/op BenchmarkVFS1TmpfsMountStat/3-6 1000000 1113 ns/op BenchmarkVFS1TmpfsMountStat/8-6 1000000 2118 ns/op BenchmarkVFS1TmpfsMountStat/64-6 100000 14251 ns/op BenchmarkVFS1TmpfsMountStat/100-6 100000 22397 ns/op BenchmarkVFS2MemfsMountStat/1-6 5000000 317 ns/op BenchmarkVFS2MemfsMountStat/2-6 5000000 361 ns/op BenchmarkVFS2MemfsMountStat/3-6 5000000 387 ns/op BenchmarkVFS2MemfsMountStat/8-6 3000000 582 ns/op BenchmarkVFS2MemfsMountStat/64-6 500000 2699 ns/op BenchmarkVFS2MemfsMountStat/100-6 300000 4133 ns/op From this we can infer that, on this machine: - Constant cost for tmpfs stat() is ~160ns in VFS2 and ~280ns in VFS1. - Per-path-component cost is ~35ns in VFS2 and ~215ns in VFS1, a difference of about 6x. - The cost of crossing a mount boundary is about 80ns in VFS2 (MemfsMountStat/1 does approximately the same amount of work as MemfsStat/2, except that it also crosses a mount boundary). This is an inescapable cost of the separate mount lookup needed to support bind mounts and mount namespaces. PiperOrigin-RevId: 258853946
2019-07-18 22:09:14 +00:00
// PRead implements FileDescriptionImpl.PRead.
func (DirectoryFileDescriptionDefaultImpl) PRead(ctx context.Context, dst usermem.IOSequence, offset int64, opts ReadOptions) (int64, error) {
return 0, syserror.EISDIR
}
// Read implements FileDescriptionImpl.Read.
func (DirectoryFileDescriptionDefaultImpl) Read(ctx context.Context, dst usermem.IOSequence, opts ReadOptions) (int64, error) {
return 0, syserror.EISDIR
}
// PWrite implements FileDescriptionImpl.PWrite.
func (DirectoryFileDescriptionDefaultImpl) PWrite(ctx context.Context, src usermem.IOSequence, offset int64, opts WriteOptions) (int64, error) {
return 0, syserror.EISDIR
}
// Write implements FileDescriptionImpl.Write.
func (DirectoryFileDescriptionDefaultImpl) Write(ctx context.Context, src usermem.IOSequence, opts WriteOptions) (int64, error) {
return 0, syserror.EISDIR
}
// DentryMetadataFileDescriptionImpl may be embedded by implementations of
// FileDescriptionImpl for which FileDescriptionOptions.UseDentryMetadata is
// true to obtain implementations of Stat and SetStat that panic.
//
// +stateify savable
type DentryMetadataFileDescriptionImpl struct{}
// Stat implements FileDescriptionImpl.Stat.
func (DentryMetadataFileDescriptionImpl) Stat(ctx context.Context, opts StatOptions) (linux.Statx, error) {
panic("illegal call to DentryMetadataFileDescriptionImpl.Stat")
}
// SetStat implements FileDescriptionImpl.SetStat.
func (DentryMetadataFileDescriptionImpl) SetStat(ctx context.Context, opts SetStatOptions) error {
panic("illegal call to DentryMetadataFileDescriptionImpl.SetStat")
}
// DynamicBytesSource represents a data source for a
// DynamicBytesFileDescriptionImpl.
//
// +stateify savable
type DynamicBytesSource interface {
// Generate writes the file's contents to buf.
Generate(ctx context.Context, buf *bytes.Buffer) error
}
// StaticData implements DynamicBytesSource over a static string.
//
// +stateify savable
type StaticData struct {
Data string
}
// Generate implements DynamicBytesSource.
func (s *StaticData) Generate(ctx context.Context, buf *bytes.Buffer) error {
buf.WriteString(s.Data)
return nil
}
// WritableDynamicBytesSource extends DynamicBytesSource to allow writes to the
// underlying source.
//
// TODO(b/179825241): Make utility for integer-based writable files.
type WritableDynamicBytesSource interface {
DynamicBytesSource
// Write sends writes to the source.
Write(ctx context.Context, src usermem.IOSequence, offset int64) (int64, error)
}
// DynamicBytesFileDescriptionImpl may be embedded by implementations of
// FileDescriptionImpl that represent read-only regular files whose contents
// are backed by a bytes.Buffer that is regenerated when necessary, consistent
// with Linux's fs/seq_file.c:single_open().
//
// If data additionally implements WritableDynamicBytesSource, writes are
// dispatched to the implementer. The source data is not automatically modified.
//
// DynamicBytesFileDescriptionImpl.SetDataSource() must be called before first
// use.
//
// +stateify savable
type DynamicBytesFileDescriptionImpl struct {
data DynamicBytesSource // immutable
mu sync.Mutex `state:"nosave"` // protects the following fields
buf bytes.Buffer `state:".([]byte)"`
off int64
lastRead int64 // offset at which the last Read, PRead, or Seek ended
}
func (fd *DynamicBytesFileDescriptionImpl) saveBuf() []byte {
return fd.buf.Bytes()
}
func (fd *DynamicBytesFileDescriptionImpl) loadBuf(p []byte) {
fd.buf.Write(p)
}
// SetDataSource must be called exactly once on fd before first use.
func (fd *DynamicBytesFileDescriptionImpl) SetDataSource(data DynamicBytesSource) {
fd.data = data
}
// Preconditions: fd.mu must be locked.
func (fd *DynamicBytesFileDescriptionImpl) preadLocked(ctx context.Context, dst usermem.IOSequence, offset int64, opts *ReadOptions) (int64, error) {
// Regenerate the buffer if it's empty, or before pread() at a new offset.
// Compare fs/seq_file.c:seq_read() => traverse().
switch {
case offset != fd.lastRead:
fd.buf.Reset()
fallthrough
case fd.buf.Len() == 0:
if err := fd.data.Generate(ctx, &fd.buf); err != nil {
fd.buf.Reset()
// fd.off is not updated in this case.
fd.lastRead = 0
return 0, err
}
}
bs := fd.buf.Bytes()
if offset >= int64(len(bs)) {
return 0, io.EOF
}
n, err := dst.CopyOut(ctx, bs[offset:])
fd.lastRead = offset + int64(n)
return int64(n), err
}
// PRead implements FileDescriptionImpl.PRead.
func (fd *DynamicBytesFileDescriptionImpl) PRead(ctx context.Context, dst usermem.IOSequence, offset int64, opts ReadOptions) (int64, error) {
fd.mu.Lock()
n, err := fd.preadLocked(ctx, dst, offset, &opts)
fd.mu.Unlock()
return n, err
}
// Read implements FileDescriptionImpl.Read.
func (fd *DynamicBytesFileDescriptionImpl) Read(ctx context.Context, dst usermem.IOSequence, opts ReadOptions) (int64, error) {
fd.mu.Lock()
n, err := fd.preadLocked(ctx, dst, fd.off, &opts)
fd.off += n
fd.mu.Unlock()
return n, err
}
// Seek implements FileDescriptionImpl.Seek.
func (fd *DynamicBytesFileDescriptionImpl) Seek(ctx context.Context, offset int64, whence int32) (int64, error) {
fd.mu.Lock()
defer fd.mu.Unlock()
switch whence {
case linux.SEEK_SET:
// Use offset as given.
case linux.SEEK_CUR:
offset += fd.off
default:
// fs/seq_file:seq_lseek() rejects SEEK_END etc.
return 0, linuxerr.EINVAL
}
if offset < 0 {
return 0, linuxerr.EINVAL
}
if offset != fd.lastRead {
// Regenerate the file's contents immediately. Compare
// fs/seq_file.c:seq_lseek() => traverse().
fd.buf.Reset()
if err := fd.data.Generate(ctx, &fd.buf); err != nil {
fd.buf.Reset()
fd.off = 0
fd.lastRead = 0
return 0, err
}
fd.lastRead = offset
}
fd.off = offset
return offset, nil
}
// Preconditions: fd.mu must be locked.
func (fd *DynamicBytesFileDescriptionImpl) pwriteLocked(ctx context.Context, src usermem.IOSequence, offset int64, opts WriteOptions) (int64, error) {
if opts.Flags&^(linux.RWF_HIPRI|linux.RWF_DSYNC|linux.RWF_SYNC) != 0 {
return 0, linuxerr.EOPNOTSUPP
}
limit, err := CheckLimit(ctx, offset, src.NumBytes())
if err != nil {
return 0, err
}
src = src.TakeFirst64(limit)
writable, ok := fd.data.(WritableDynamicBytesSource)
if !ok {
return 0, syserror.EIO
}
n, err := writable.Write(ctx, src, offset)
if err != nil {
return 0, err
}
// Invalidate cached data that might exist prior to this call.
fd.buf.Reset()
return n, nil
}
// PWrite implements FileDescriptionImpl.PWrite.
func (fd *DynamicBytesFileDescriptionImpl) PWrite(ctx context.Context, src usermem.IOSequence, offset int64, opts WriteOptions) (int64, error) {
fd.mu.Lock()
n, err := fd.pwriteLocked(ctx, src, offset, opts)
fd.mu.Unlock()
return n, err
}
// Write implements FileDescriptionImpl.Write.
func (fd *DynamicBytesFileDescriptionImpl) Write(ctx context.Context, src usermem.IOSequence, opts WriteOptions) (int64, error) {
fd.mu.Lock()
n, err := fd.pwriteLocked(ctx, src, fd.off, opts)
fd.off += n
fd.mu.Unlock()
return n, err
}
// GenericConfigureMMap may be used by most implementations of
// FileDescriptionImpl.ConfigureMMap.
func GenericConfigureMMap(fd *FileDescription, m memmap.Mappable, opts *memmap.MMapOpts) error {
opts.Mappable = m
opts.MappingIdentity = fd
fd.IncRef()
return nil
}
// LockFD may be used by most implementations of FileDescriptionImpl.Lock*
// functions. Caller must call Init().
//
// +stateify savable
type LockFD struct {
locks *FileLocks
}
// SupportsLocks implements FileDescriptionImpl.SupportsLocks.
func (LockFD) SupportsLocks() bool {
return true
}
// Init initializes fd with FileLocks to use.
func (fd *LockFD) Init(locks *FileLocks) {
fd.locks = locks
}
// Locks returns the locks associated with this file.
func (fd *LockFD) Locks() *FileLocks {
return fd.locks
}
// LockBSD implements FileDescriptionImpl.LockBSD.
func (fd *LockFD) LockBSD(ctx context.Context, uid fslock.UniqueID, ownerPID int32, t fslock.LockType, block fslock.Blocker) error {
return fd.locks.LockBSD(ctx, uid, ownerPID, t, block)
}
// UnlockBSD implements FileDescriptionImpl.UnlockBSD.
func (fd *LockFD) UnlockBSD(ctx context.Context, uid fslock.UniqueID) error {
fd.locks.UnlockBSD(uid)
return nil
}
// LockPOSIX implements FileDescriptionImpl.LockPOSIX.
func (fd *LockFD) LockPOSIX(ctx context.Context, uid fslock.UniqueID, ownerPID int32, t fslock.LockType, r fslock.LockRange, block fslock.Blocker) error {
return fd.locks.LockPOSIX(ctx, uid, ownerPID, t, r, block)
}
// UnlockPOSIX implements FileDescriptionImpl.UnlockPOSIX.
func (fd *LockFD) UnlockPOSIX(ctx context.Context, uid fslock.UniqueID, r fslock.LockRange) error {
return fd.locks.UnlockPOSIX(ctx, uid, r)
}
// TestPOSIX implements FileDescriptionImpl.TestPOSIX.
func (fd *LockFD) TestPOSIX(ctx context.Context, uid fslock.UniqueID, t fslock.LockType, r fslock.LockRange) (linux.Flock, error) {
return fd.locks.TestPOSIX(ctx, uid, t, r)
}
// NoLockFD implements Lock*/Unlock* portion of FileDescriptionImpl interface
// returning ENOLCK.
//
// +stateify savable
type NoLockFD struct{}
// SupportsLocks implements FileDescriptionImpl.SupportsLocks.
func (NoLockFD) SupportsLocks() bool {
return false
}
// LockBSD implements FileDescriptionImpl.LockBSD.
func (NoLockFD) LockBSD(ctx context.Context, uid fslock.UniqueID, ownerPID int32, t fslock.LockType, block fslock.Blocker) error {
return linuxerr.ENOLCK
}
// UnlockBSD implements FileDescriptionImpl.UnlockBSD.
func (NoLockFD) UnlockBSD(ctx context.Context, uid fslock.UniqueID) error {
return linuxerr.ENOLCK
}
// LockPOSIX implements FileDescriptionImpl.LockPOSIX.
func (NoLockFD) LockPOSIX(ctx context.Context, uid fslock.UniqueID, ownerPID int32, t fslock.LockType, r fslock.LockRange, block fslock.Blocker) error {
return linuxerr.ENOLCK
}
// UnlockPOSIX implements FileDescriptionImpl.UnlockPOSIX.
func (NoLockFD) UnlockPOSIX(ctx context.Context, uid fslock.UniqueID, r fslock.LockRange) error {
return linuxerr.ENOLCK
}
// TestPOSIX implements FileDescriptionImpl.TestPOSIX.
func (NoLockFD) TestPOSIX(ctx context.Context, uid fslock.UniqueID, t fslock.LockType, r fslock.LockRange) (linux.Flock, error) {
return linux.Flock{}, linuxerr.ENOLCK
}
// BadLockFD implements Lock*/Unlock* portion of FileDescriptionImpl interface
// returning EBADF.
//
// +stateify savable
type BadLockFD struct{}
// SupportsLocks implements FileDescriptionImpl.SupportsLocks.
func (BadLockFD) SupportsLocks() bool {
return false
}
// LockBSD implements FileDescriptionImpl.LockBSD.
func (BadLockFD) LockBSD(ctx context.Context, uid fslock.UniqueID, ownerPID int32, t fslock.LockType, block fslock.Blocker) error {
return linuxerr.EBADF
}
// UnlockBSD implements FileDescriptionImpl.UnlockBSD.
func (BadLockFD) UnlockBSD(ctx context.Context, uid fslock.UniqueID) error {
return linuxerr.EBADF
}
// LockPOSIX implements FileDescriptionImpl.LockPOSIX.
func (BadLockFD) LockPOSIX(ctx context.Context, uid fslock.UniqueID, ownerPID int32, t fslock.LockType, r fslock.LockRange, block fslock.Blocker) error {
return linuxerr.EBADF
}
// UnlockPOSIX implements FileDescriptionImpl.UnlockPOSIX.
func (BadLockFD) UnlockPOSIX(ctx context.Context, uid fslock.UniqueID, r fslock.LockRange) error {
return linuxerr.EBADF
}
// TestPOSIX implements FileDescriptionImpl.TestPOSIX.
func (BadLockFD) TestPOSIX(ctx context.Context, uid fslock.UniqueID, t fslock.LockType, r fslock.LockRange) (linux.Flock, error) {
return linux.Flock{}, linuxerr.EBADF
}